The Costs of Seeking Certainty in the Midst of a Pandemic
Should we wait for more certainty about COVID-19 before taking drastic measures?
In a recent piece called “A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data”, Stanford researcher John Ioannadis has suggested that we should pump the brakes on drastic measures like lockdowns until we have more certainty about the course of COVID-19. Cognitive psychologist Gerd Gigerenzer has suggested that we might be overreacting to the pandemic in the same way that we overreact to the vivid-but-uncommon threat of terrorism. What if COVID-19 turns out to be less deadly than the flu? Won’t we regret the damage to the economy and the panic that’s been induced?
These are reasonable questions, at an unreasonable time. What these takes are missing, though, is that there is a cost to seeking certainty — and in the case of COVID-19 that cost in measured in lives.
In the science of decision making, it is axiomatic that there will frequently be trade-offs between time and accuracy. Generally, the more time we take to gather information before deciding, the more accurate our decisions will be. The less time we take, less accurate. It’s a no-brainer that we would make better decisions if we had better information; and we will surely have better information about COVID-19 a week from now, and still better information a month or a year from now.
But that doesn’t mean we should wait that week or month or year before taking action. Another important principle of decision making is that waiting for more information comes with a cost. The lower the cost to waiting, the more we can afford to wait while gathering more data; the higher the cost, the less can afford to wait before taking action. In an ideal world, we would wait for complete information before deciding to take radical measures that would largely shut the economy down. But in the real world as this crisis is unfolding, we don’t have that luxury.
If the course the pandemic takes were the same whether we acted now or in three weeks, waiting for more certainty might make sense. The time it would take to gather more data might not matter. But that is not how a virus that grows exponentially works. Every day matters for how many people get sick and how many people die.
In countries that responded swiftly and firmly to the virus, like South Korea and Taiwan, the rapid increase in COVID-19 has been largely attenuated. In places like Italy and The United States that have been slow to react, the disease has spread wildly. In the U.S. alone, the number of deaths has more than doubled in the last three days; over the last two weeks the number of confirmed cases has increased over fifty-fold.
Furthermore, as has become clear, with equipment shortages and limits on hospital capacity, mortality rates increase rapidly as a function of the number of cases relative to the available resources. In countries like South Korea where they didn’t miss the containment window, the case fatality rate appears to be hovering around .6%. In a country like Italy, where policy makers waited for more data before acting, the case fatality rate appears to be somewhere between 3% and 4%. We expect to see New Zealand, which locked down after 120 confirmed cases, weather the storm far better than places like Italy, Canada, and the United States, that have waited for more data, instituting fewer lock downs when the disease has already spread more widely.
The cost in lives to waiting is undeniable.
Rather than waiting to buy certainty that we can ill afford, we should be trying to buy time. Because of the exponential nature of the rate of infection, time is the factor that can most help us. An analysis from Tomas Pueyo suggests that acting just one day earlier in the decision to lockdown might result in a 40% reduction in cases.
The other problem with dallying is that mitigation strategies often become more expensive the longer you wait. If you have just built a brand-new house and you want to buy fire insurance, it isn’t going to cost you much. But if you wait until there is a raging wildfire heading in your direction, fire insurance is now going to be really expensive. Yes, you have much more data about whether your house is going to burn down. But the wait for certainty is costly.
In January as the virus was raging in China, there was some chance it would hit the U.S. hard. The worst-case scenarios suggested that millions of Americans could die. The best-case scenarios suggested it might never get a foothold here and, even if it did, it might be no worse than the flu. The course the virus would take here was very uncertain. But in the months of January and February, exactly when uncertainty was high, the U.S. could have been manufacturing and stockpiling tests. We could have been building ventilators and making masks. Compared to the national debt, those hedges would have been practically free. The cost of that insurance would have been small; fixing things now is going to be vastly more expensive.
We buy insurance hoping we never have to use it. Our leaders chose not to buy the insurance, perhaps in part because they were concerned about the costs in case the virus petered out. But the costs of waiting for more certainty have been catastrophic. Let’s not keep repeating the same deadly cognitive error. Let’s not continue to allow inaction based on uncertainty lead to further exponential losses.
So what can we do now? We can’t turn back the clock and restore the time that could have been so valuable. But we can work with the information we’ve got, factoring in the staggering costs of waiting.
We already know, with virtual certainty, that the virus spreads exponentially when left unchecked; we know that measures that seemed draconian to Western civilizations a few weeks ago don’t suffice; we know that few countries if any have enough tests for all their citizens or enough hospital beds.
As countries like France, Italy, and, most recently, Britain, India, New Zealnd, and South Africa have chosen to follow the lead of Wuhan and implement nationwide lockdowns, they are buying themselves time. If the U.S. and Canada follows their lead, North America can can buy time, too. We can use that time to start producing tests at a scale that allows us to test virtually everybody. We can do the science we need, so we can figure out how much asymptomatic transmission matters. We can let our hospitals catch up, restock their supplies of masks, ventilators, and beds.
Eventually, if we buy enough time, we can position ourselves to follow the test-and-trace strategy that has been so effective in South Korea, and the Italian town of Vo (which tested all 3300 of its residents and has successfully controlled COVID-19 within its territory).
We can do all that here — if the public and our leaders have the will.
Annie Duke is author of Thinking in Bets, and co-founder of The Alliance for Decision Education
Gary Marcus, PhD, is a scientist and entrepreneur, and Professor Emeritus at New York University; @garymarcus on Twitter.
